I was saddened to see the videos and read the news stories about the flooding and landslides which occurred in Japan over the past several days.

Photo: Baltimore Sun

As of this writing, more than 179 people have been confirmed killed and 8 million people have been ordered to evacuate their homes.

Photo: Toronto Star

The flooding in Japan illustrates that Houston is not alone in dealing with flood risk. The news prompted me to do some research into the situation in Japan. This post provides some additional information about flood risk management and hydrology in Japan.

The first important difference between Japan and Houston is topography.  The rivers in Japan are very steep and convey water with high velocities.  The graphic below, which I adapted from a figure in an un-dated publication of the Network of Asian River Basin Organizations (NARBO) called “Flood Management in Japan,” illustrates the river profiles (horizontal distance and vertical fall) for selected rivers in Japan (blue) and some selected rivers in other countries (green).

When Dutch engineer Johannis de Rijke saw the Jōganji River in Toyama Prefecture (the steepest one shown above – to the far left) he famously said: “this is not a river [its] a waterfall.”

The topography in Houston is very, very flat. I added the approximate profile of Buffalo Bayou from the Barker Dam to Upper Galveston Bay in pink.

Flood risk literature generally reports that about 5% to 14% of the population of the United States lives in a mapped flood risk zone (1% annual chance).  In Japan, due to the steep, mountainous terrain, the populated areas are located in the flatter, alluvial plains between mountains and towards the mouths of most rivers.  The publication referenced above indicates that 50% of the Japanese population are located in these alluvial plains. I was not able to locate flood risk mapping products indicating what annual chance risk exposure they have.

I was able to locate information about rainfall statistics. A paper from 2017 by Sugi, et. al. used a statistical method called “extended regional frequency analysis” (ERFA) to evaluate three sets of rainfall data to estimate the depth of the 1% annual chance, 24-hour storm. This method is similar to the Atlas 14 project approach I’ve posted about on September 27, 2017, November 20, 2017, November 24, 2017, December 7, 2017, and February 21, 2018.  The three sets of rainfall data they evaluated were:

• Actual rainfall observations from 1979 to 2006, across a wide (regional) geographical area;
• Modeled rainfall observations using ocean temperatures, atmospheric carbon dioxide concentrations, and atmospheric aerosol concentrations from 1979 to 2003, across the same wide geographical area; and,
• Modeled rainfall observations using ocean temperatures, atmospheric carbon dioxide concentrations, and atmospheric aerosol concentrations predicted from the IPCC RCP8.5 scenario for 2075 to 2099, across the same wide geographical area. The RCP8.5 is a “worst case” simulation of unrestricted carbon emissions.

The results of these evaluations are presented below in a graphic that I adapted from the paper.

The key takeaways:

• Some of the inland, higher elevation areas of Japan have a less than 1% annual chance of getting 10 or more inches of rain in 24-hours.
• Some coastal areas in the south have a 1% annual chance of seeing 24 inches or more of rain in 24-hours. Compare that to Houston’s current 1% annual chance, 24-hour rain depth of about 13.5 inches.
• Climate change is projected to expand the areas with a 1% annual chance of seeing 24 inches of rain, or more in 24-hours.

Now let’s compare these data to what happened in Japan on July 6, 2018. Much like our own Harris County Flood Control District “Flood Warning System” gauge network (with 163 gauges), the Japan Meteorological Agency (JMA) operates a network of 160 rain gauges and 20 Doppler radar stations across Japan. On July 6, 2018, the JMA recorded rainfall depths up to 15 inches in 24-hours in specific locations and over 7 or 8 inches of rain over widespread areas of southern Japan. The graphic below was adapted from JMA historical data maps obtained from their data web page.

The gauge locations with the black diagonal line and box indicate new 24-hour rainfall depth total records for that gauge location.

So the high rainfall totals coupled with the steep terrain and people and infrastructure located in the alluvial plains all combined to lead to the loss of life and infrastructure damage.